1.Claim Autopsy - Evidence Analysis AssistantEnter a claim you want checked and it separates the main claim from its smaller claims, hidden assumptions, and the evidence needed, then looks at the evidence and gives a final verdict.Using AI › Hallucination checks · #Analyzing #Reviewing2.FactcheckEnter text to check to have its verifiable claims pulled out and sorted into confirmed, needs-recheck, and false-or-unsupported, with suggested corrections.Using AI › Hallucination checks · #Reviewing #Analyzing3.Fact-Checking Evaluation AssistantAdd a claim and source excerpts to get the key points pulled out, source reliability judged, support checked, the opposing view argued back, and a final verdict with brief reasoning.Using AI › Hallucination checks · #Analyzing #Reviewing #Checklist4.🧠 FORMAL VERIFICATION MODEAdd a claim or problem to verify and get an answer laid out as givens, assumptions, step-by-step reasoning, a consistency check, and a final judgment with a confidence level.Using AI › Hallucination checks · #Analyzing #Reviewing #Checklist5.Catch fabricated citationsWorks from the premise that invented citations look more correct than real ones.Using AI › Hallucination checks · #Reviewing #Analyzing #Table6.Make an answer state its own certaintyLabels every statement established, likely, inferred, or unknown — and defaults downward when unsure.Using AI › Hallucination checks · #Reviewing #Rewriting #Analyzing7.Read what a disagreement between models meansTreats agreement as no proof of correctness and flags a claim only one model made.Using AI › Hallucination checks · #Analyzing #Reviewing #Table8.Pick out what in an AI answer needs verifyingSorts claims by what goes wrong if they are false, and names where to check each one.Using AI › Hallucination checks · #Reviewing #Analyzing #Checklist9.Make the model say it does not knowMakes a blank an acceptable answer, and makes a false premise something to name rather than answer around.Using AI › Hallucination checks · #Question generation #Analyzing #Reviewing10.Hallucination Vulnerability Prompt CheckerPaste in a prompt to review and it finds weak spots that can lead an AI to make things up, such as forced answers, requests for unsupported data, vague generalizations, and drifting off topic, then explains each with a fix.Using AI › Hallucination checks · #Analyzing #Reviewing #Checklist11.Separate calculated numbers from recalled onesFlags figures the model produced from training, and treats suspicious precision as a warning.Using AI › Hallucination checks · #Reviewing #Analyzing #Table12.Idea Reality Check - Am I Actually Onto Something?Share a new idea, invention, theory, or observation and it checks, one step at a time, the core claim and assumptions, how it compares with existing ideas, and whether it could actually work.Using AI › Hallucination checks · #Analyzing #Reviewing13.Large Language Models Security SpecialistEnter example prompts meant to test an LLM and get mitigation methods and security guidance for risks like data exposure, prompt injection, and harmful content.Using AI › Hallucination checks · #Analyzing #Reviewing #Question generation14.Plan Check AgentEnter a strategy or plan and it keeps hunting for shaky assumptions, gaps, and fixes, checking the plan over and over until it's confident it holds up.Using AI › Hallucination checks · #Reviewing #Analyzing #Checklist15.Potato CriticType an idea or claim after 'Potato' to get 3 logical failure points, 2 unproven assumptions, and 1 missing counterargument flagged.Using AI › Hallucination checks · #Reviewing #Analyzing16.🛡️ RED TEAM MODEEnter an idea, plan, claim, or system and get an aggressive analysis of hidden assumptions, weaknesses, failure scenarios, impact, and fixes.Using AI › Hallucination checks · #Analyzing #Reviewing #Checklist17.Semantic Prosody–Based Epistemic Bias Correction PromptWhen working on a question or draft answer, it checks whether key words are quietly framing the issue around one field, culture, or value system, and offers alternatives.Using AI › Hallucination checks · #Reviewing #Analyzing18.Western-Centric Bias Correction PromptAsk a question and it answers without treating Western societies as the default, widening the context, sources, people involved, and historical background.Using AI › Hallucination checks · #Reviewing #Analyzing
TukttakAI prompt library